The 2017-18 Premier League season may go down as one of the most significant in football history. Manchester City, led by legendary manager Pep Guardiola, broke Premier League records for biggest winning margin, most consecutive wins, most points, most goals, and best goal difference. This level of sustained brilliance was unprecedented and set a new benchmark for top-flight success.
However, the introduction of the expected goals (xG) metric on BBC’s Match of the Day programme that season was arguably even more revolutionary for the footballing world. Statistical data was included in the mainstream football conversation, at a time when it was already a staple of American sports. In the US, the analytics revolution had started with ‘Moneyball’ in Major League Baseball (MLB), but has since become a key part of the other Big Four sports, including the National Basketball Association (NBA).
Following the installation of player-tracking cameras in all NBA arenas in 2013, basketball has incorporated data analysis into the decision-making process for resting players, recruitment, and in-game coaching. For instance, testing players’ saliva became commonplace as a way to gauge their fatigue levels. Shot maps, which showed the percentage of shots that were made, out of those taken, from each area of the court were used by coaches to determine which shots were most efficient. One of the revolutionary changes in the way basketball was played, the increased reliance on the three-point shot, owed much to calculations of the expected points metric from different areas of the floor.
The recent stats revolution of football has coincided with the rise of data-collection services such as Opta Sports. Liverpool FC, the Premier League leaders, have famously taken a statistics-based approach to transfers, as their American owners have encouraged the use of data in transfer decisions; the club’s owner, John W. Henry, is also the owner of the Boston Red Sox, an MLB team, and has brought the Moneyball mentality across the pond. Much like in the NBA, data analysis has not just been used for recruitment but has also caused an evolution in the way football is played. Crossing, once a prominent tactic in English football, has been shunned in favour of ground passes as the data has shown the latter to be more effective in chance creation. Additionally, the popularity of long-range shooting has declined, since long shots are unlikely to result in a goal, even for specialists in the art like Ruben Neves.
Sporting data, however, is not homogeneous across different sports. Kirk Goldsberry, who is credited with leading the NBA’s statistical revolution, draws a clear distinction between the probabilistic data found in baseball and the cartographical data in basketball. While stop-start sports, such as baseball, are better-suited to analytics, dynamic sports, such as football or basketball, are better approached in a different way. Since spatial positioning is as integral to football as it is to basketball, football has much to learn from the NBA’s usage of cartography, such as shot maps. Premier League teams could use empirical data analysis to determine the zones from which each player shoots most accurately and incorporate the results into their tactical plan. Given that out-of-context discrete data, such as assist counts, is often useless for making qualitative judgements, in both football and basketball, it is time for the Premier League to follow the NBA’s reliance on contextual data. For example, football could place a heavier emphasis on efficiency over raw numbers in the case of goals, by looking at the deviation from xG when judging a striker’s effectiveness. The use of data analysis in football may be more difficult, given how the passages of play are not as easily segmented as in the MLB or NBA, but many insights from the MLB or NBA world can still be implemented in the sport.
Although data analysis in football may never reach the level found in the MLB, NBA or National Football League (NFL) due to its more fluid nature, football should seek to close the gap. Given the fine margins that make the difference in the sport, insights found in the data could well convert a loss into a key win.
The rise of medical science in football in recent years has allowed players to stay at peak fitness throughout the season. Now it is time to build on that scientific progress by placing greater emphasis on in-game data analysis.